What is an ETF? In simple terms, an exchange-traded fund is an investment fund whose shares can be bought and sold on a stock exchange much like ordinary company shares. Behind that simple definition, however, sits a more complex structure involving the fund itself, its underlying assets, market makers, authorized participants and a mechanism designed to keep the ETF’s market price close to the value of the assets it holds.
Understanding how these pieces fit together is important because an ETF is not simply a basket of shares displayed inside a brokerage account. When an investor buys an ETF, they acquire shares in the fund itself. The fund then owns the underlying securities, which can include stocks, bonds, commodities or other financial assets depending on its strategy. The U.S. Securities and Exchange Commission describes ETFs as pooled investment vehicles whose shares trade on national stock exchanges.
For example, an ETF tracking the S&P 500 may own shares in hundreds of large US companies in proportions designed to resemble the index. An investor who buys that ETF gains economic exposure to those companies, but does not directly own small positions in each one of them. Legally, the investor owns shares in the ETF, while the ETF owns the securities in its portfolio.
What is an ETF – How an ETF gets its value
One of the most important concepts for understanding ETFs is net asset value, usually referred to as NAV. It represents the total value of the fund’s assets after liabilities are deducted, divided by the number of ETF shares outstanding.
If a fund owns assets worth $1.001 billion, has liabilities of $1 million and has 10 million ETF shares outstanding, its NAV would be $100 per share.
This does not necessarily mean the ETF will trade at exactly $100 on the stock exchange. Unlike a traditional mutual fund, where investors generally transact at a calculated end-of-day NAV, ETF shares trade continuously during market hours. Their price is therefore determined by supply and demand.
An ETF with a NAV of $100 might trade at $100.10 if buyers are particularly eager to acquire it. In that case, the ETF trades at a premium to NAV. If its market price falls to $99.90 while the portfolio remains worth $100 per share, it trades at a discount.
For large and liquid ETFs, these differences are usually small. An important part of the ETF structure is specifically designed to prevent the market price from moving too far away from the value of the underlying portfolio.
Read also: Analysts: Main Investment Trends This Year Include ETF Purchases and AI Utilization
How creation and redemption keep ETF prices in line
Retail investors generally buy and sell ETF shares on an exchange. They usually have no direct relationship with the fund provider. Behind those everyday trades, however, large financial institutions known as authorized participants play a crucial role.
Authorized participants can create new ETF shares or redeem existing ones in large blocks. This process links the ETF shares traded on the stock market to the securities held by the fund.
Suppose an ETF tracking a stock index becomes very popular and its market price rises to $101 even though the underlying portfolio is worth only $100 per ETF share. An authorized participant may be able to buy the required underlying securities for approximately $100, deliver them to the ETF provider and receive newly created ETF shares in return.
Those ETF shares can then be sold on the market for around $101. The process increases the supply of ETF shares and creates pressure that can bring the market price closer to NAV.
The reverse can happen if the ETF trades below the value of its underlying assets. If an ETF with a NAV of $100 trades at $99, an authorized participant may buy ETF shares on the market, return them to the fund and receive underlying securities worth approximately $100. This reduces the number of ETF shares available and can help move the price back toward the value of the portfolio.
This mechanism, known as creation and redemption, is one of the defining features of the ETF structure. The Investment Company Institute explains that authorized participants exchange large creation units directly with the fund, typically using baskets of securities, cash or a combination of both.
The system is not perfect. During periods of severe market volatility, when underlying assets are illiquid or when the markets in which those assets trade are closed, premiums and discounts can become larger. Under normal market conditions, however, the mechanism generally helps keep the two prices relatively close.
What is an ETF – Why the price you pay is not exactly NAV
Another cost investors encounter is the bid-ask spread.
Every exchange-traded security normally has a bid price, representing the highest price buyers are currently willing to pay, and an ask price, representing the lowest price sellers are willing to accept.
Imagine an ETF is quoted at $99.98 bid and $100.02 ask. An investor buying the fund will normally pay around $100.02. If they sold it immediately, they might receive around $99.98.
The difference of four cents represents the bid-ask spread.
It may seem insignificant, particularly for a large and highly liquid ETF, but it is still a real trading cost. Unlike an annual fund fee, it will not normally appear as a separate charge on an investor’s statement. Instead, it is embedded in the price at which the transaction is completed.
Spreads can vary substantially between funds. Large ETFs investing in highly liquid stocks often trade with very narrow spreads. More specialized ETFs holding smaller companies, emerging-market securities or less liquid bonds may have wider spreads because market makers face greater costs and risks when trading or hedging the underlying assets.
Market conditions also matter. Even normally liquid ETFs can experience wider spreads during periods of unusually high volatility.
What the ETF expense ratio actually means
Investors also need to consider the ETF’s expense ratio. This represents the annual operating costs of the fund as a percentage of its assets.
An ETF with an expense ratio of 0.20% effectively costs around $20 per year for every $10,000 invested, assuming the investment remains around that value.
The investor does not receive a $20 bill from the ETF provider. Instead, the cost is deducted from the fund’s assets over time. It therefore gradually affects the ETF’s net asset value and ultimately the return received by investors.
This is one reason why an ETF tracking an index will not normally deliver exactly the same return as the index itself.
Why an ETF never tracks an index perfectly
An index is essentially a mathematical benchmark. An ETF, by contrast, is a real investment portfolio that has to buy, sell and hold actual securities.
Some ETFs use full replication, meaning they attempt to own all the securities included in their benchmark in roughly the same proportions. Others use sampling, holding only part of the index while attempting to reproduce its overall behaviour. Depending on the type of ETF, derivatives can also be used.
In practice, several factors can cause returns to deviate from the benchmark.
The fund has operating expenses. It may also incur transaction costs when the index is rebalanced. It can temporarily hold cash from dividends or investor flows. Foreign taxes can affect distributions, while differences in valuation times can create additional discrepancies.
This leads to two concepts that are often confused: tracking difference and tracking error.
Tracking difference describes the gap between the ETF’s return and the return of its benchmark over a particular period. If an index rises by 10% but an ETF rises by only 9.8%, the tracking difference is 0.2 percentage points.
Tracking error looks at how consistently the ETF follows the benchmark over time. An ETF may lag its index by a small and predictable amount every year because of its fees while still having a relatively low tracking error. Another ETF may sometimes outperform the index and sometimes lag significantly, creating a higher tracking error even if its average long-term difference is similar.
As Vanguard notes, investors should therefore consider not only the expense ratio, but also trading costs and how efficiently the ETF tracks its benchmark.
For investors comparing two ETFs following the same benchmark, this can be more useful than simply looking at which fund advertises the lowest annual fee.
Read also: Best Sectors to Invest in 2026: 7 Industries Positioned for Structural Growth
What happens when you invest $10,000 in an ETF?
A practical example makes the mechanics easier to understand.
Imagine an ETF has a NAV of exactly $100 per share. The market currently quotes a bid price of $99.98 and an ask price of $100.02. The fund has an annual expense ratio of 0.20%.
An investor decides to put $10,000 into the ETF.
Assuming the broker allows fractional shares, buying at the $100.02 ask price would give the investor approximately 99.98 ETF shares.
Those shares are what appear in the brokerage account. The investor does not suddenly receive hundreds of separate positions in every company held by the fund.
At a NAV of $100, the investor’s 99.98 shares represent around $9,998 of underlying net assets. The small difference compared with the original $10,000 reflects the fact that the ETF was purchased slightly above NAV because of the spread.
Suppose the ETF tracks an index containing 500 companies and one of those companies represents 7% of the fund. The investor would have roughly $700 of economic exposure to that company through the ETF.
But the investor does not directly own $700 of that company’s shares. The ETF owns the shares. The investor owns part of the ETF.
Now imagine the underlying index rises by 8% over the next year.
If there were no costs or tracking differences, the $9,998 of underlying exposure would increase to around $10,798.
In reality, the ETF charges an annual expense ratio of 0.20%. Around $20 of annual costs would therefore be associated with a position of approximately $10,000, although this amount is reflected in the fund’s performance rather than billed separately.
Suppose the fund also loses another 0.05 percentage points relative to the benchmark because of trading costs, cash holdings and other portfolio-management effects. Instead of delivering the full 8% index return, the ETF might return around 7.75%.
The position would then be worth approximately $10,773 before taxes and before any costs associated with eventually selling the ETF.
This simple example shows where ETF costs actually appear. Some are visible in the expense ratio. Others are embedded in trading prices or appear as a small difference between the performance of the ETF and its benchmark.
A broker may also charge commissions, currency-conversion costs or other fees. Tax treatment will depend on the investor’s country of residence and the structure and domicile of the ETF itself.
An ETF and an index are not the same thing
It is therefore useful to distinguish between an index and the ETF that tracks it.
An index is a benchmark showing how a theoretical group of securities performs according to a specific methodology. An ETF is an investable financial product that attempts to reproduce that performance in the real world.
That difference matters particularly when comparing similar ETFs.
Two funds may both track the same index but produce slightly different returns because they have different fees, replication methods, tax treatment, trading costs or securities-lending policies.
Liquidity can differ as well. A large ETF with billions of dollars in assets may trade much more efficiently than a small fund tracking the same benchmark.
The fund’s domicile can also be important, particularly for international investors, because taxation of dividends and capital gains may vary between jurisdictions.
What is an ETF and why does the structure matter?
The question what is an ETF? therefore has a more precise answer than simply describing it as a basket of investments.
An ETF is a fund whose shares trade on an exchange. Investors own shares in that fund, while the fund owns the underlying assets. Its net asset value reflects the value of those assets, while its market price changes throughout the trading day.
Authorized participants and the creation-redemption mechanism help connect those two values and prevent large discrepancies under normal market conditions.
For investors, the most important costs are not limited to the advertised expense ratio. Bid-ask spreads, tracking differences, brokerage charges, currency conversion and taxes can all affect the final return.
Understanding those mechanics does not make ETFs more complicated investments. It simply makes clear what investors are actually buying when they place an ETF order — and where the gap between an index return and the return in their own portfolio can come from.
